Thinklytics

Austin Community College · Higher Education · Austin, TX · 14 weeks

Student success analytics deployed to 220 advisors

Austin Community College relied on a manual process to identify at-risk students, which took three weeks each semester. This delay prevented timely intervention. We developed a student success analytics platform that flagged at-risk students within 48 hours of early warning signs, giving advisors enough time to act before students dropped out.

Challenge

The advising team spent three weeks each semester manually gathering enrollment, grade, and attendance data from three different systems and compiling it in Excel. This delayed identifying at-risk students, leaving just two to three weeks for interventions before the withdrawal deadline. As a result, the college lost between 1,800 and 2,200 students each semester to avoidable withdrawals.

Approach

We created a Power BI platform that pulls daily data from three systems to calculate an at-risk score for each student using 18 indicators. The platform highlights the highest-risk students in a dashboard for advisors to prioritize their caseloads. We trained 220 advisors on how to use it over three weeks.

Outcome

We cut the time to identify at-risk students from three weeks to two days. In the first semester using the new system, advisors reached out to 1,240 students who were likely to drop out based on past data. This led to a 4.2-point increase in retention, securing $3.1 million in tuition that otherwise would have been lost.

18 Warning Signs We Caught Before Things Went Sideways

Here’s what we did: we built an at-risk score based on 18 different factors that show how students hang in there at community college. Things like missed assignments, grade trends, attendance, whether they got financial aid, shifts in their course load, and how they performed the previous semester. When we combined all these pieces, it gave us a way clearer picture of who was likely to drop out.

We spent our advisor time where it actually made a difference.

We built the platform to flag students who were high-risk. Then, we paired those students with advisors who knew just what to do. That way, advisors spent their time where it counted most, helping students who needed urgent support.

How We Kept 1,240 Students Engaged All Semester Long

We jumped in and helped the college bump up first-semester retention by 4.2 points. That might sound small, but it meant 1,240 more students didn’t drop out. With tuition at $2,500 a semester, that’s an extra $3.1 million in the bank. Frankly, the platform basically paid for itself in just one semester.

Results

  • 1,240 Students retained in first semester
  • 3 wks to 48 hrs At-risk identification time
  • $3.1M Retained tuition revenue
  • 220 Advisors using the platform

Before, we only identified at-risk students about three weeks into the semester, so we didn’t have much time to intervene. Now, we catch them within 48 hours. As a result, we prevented 1,240 students from dropping out in the first semester. Those are 1,240 students who stayed enrolled.

Vice President of Student Affairs, Austin Community College

In the client's terms

The question

What decisions is this reporting supposed to support?

The situation

Each semester the advising team spent three weeks pulling enrolment, grade and attendance data from three systems into Excel. That left two to three weeks to intervene before the withdrawal deadline, and the college was losing 1,800 to 2,200 students a semester to avoidable withdrawals.

What we found

The intervention worked when it happened. The problem was that the list of who needed it arrived with almost no time left to act, so the constraint was identification speed rather than advising capacity.

What it produced

Identification fell from three weeks to two days. In the first semester advisers reached 1,240 students the model flagged as likely to withdraw. Retention rose 4.2 points, holding $3.1M in tuition.

The roadmap

A Power BI platform pulling daily from all three systems and scoring every student against 18 risk indicators. Advisers see their caseload ranked rather than a report they have to interpret. Two hundred and twenty advisers were trained over three weeks, because a ranked list nobody trusts gets ignored.

The work behind this result

Related case studies

Frequently asked questions

How do you identify at-risk students early enough to intervene?

Score continuously rather than at census. Austin Community College cut identification from three weeks to two days using 18 indicators including missed assignments, grade trajectory, attendance, financial aid status and course load changes.

What is student retention worth?

At Austin Community College, 4.2 points of first-semester retention was 1,240 students and $3.1M in tuition held, which covered the platform inside one term.

Why do advisers not act on at-risk reports?

Usually because the list arrives too late, not because they disagree with it. Here the manual process consumed three of the five weeks available before the withdrawal deadline. Advisers now see their own caseload ranked rather than a report they have to interpret.